Please think deeply and advise from the following perspectives
(The one I mistakenly assumed now is the 18th)
Below is a brief recap of last week (3/12 - 3/18) and our future plans.
.
(hereinafter abbreviated)
English translation system repair (T0036).
Scrapbox automatic English translation function was restored.
Plurality Tokyo Namera Conference open to the public on YouTube (T0035, T0068).
Video release completed on 3/6.
Convert SmartNews Institute books to PDF and have AI read them (T0096).
Completed on 3/8.
During the past week, we have been able to complete several larger tasks, including video release and English translation system repair.
--
.
1.### 3/9 Prepare materials for LT presentation (T0067)
.
- Top priority since the date is tomorrow (3/9).
- Probably a situation of lack of time and other task priorities that have not been taken care of.
- It's done.
2.### Tax return (T0109) . - The deadline is 3/15 and there is less than a week left. - Without detailed breakdowns, it is easy to make last-minute progress.
3.### MVID dataset analysis (e.g., T0008) . - Dependent tasks (T0009, T0010, etc.) are also stopped. - The lack of clear deadlines tends to put them off.
4.### AI Instance Role Definition Document (T0012) . - Dependent source T0011 has been completed but not yet started. - Documentation tasks tend to have a lower priority for initiation. - I thought I cancelled this one ambiguously...
5.### Unexplored night-related preparations . - The task appears to be in the same vein as "3/9 LT (T0067)" and has yet to be documented.
6.### Measures to deal with increased information channels (T0113 / T0116) . - There are duplicate tasks with the same purpose (T0113, T0116), but they have not been initiated. - The problem of too many Slack/Signal/ChatGPT, etc. and confusion has taken a back seat.
.
--
.
. - Next important: deadline is 3/15, so we need to finish the general framework by the end of the weekend. - Breakdown Example - Identification and preparation of necessary materials - Summary of Income and Expenses - e-Tax input/submission procedure check
2.### AI Instance Role Definition (T0012) → Task Management Knowledge Documentation (T0013) - T0011 (task recording mechanism) is finished and can be moved. - The overall AI coordination is becoming more complex, so organizing roles now will reduce confusion.
3.### MVID dataset analysis (T0008) . - Not pressed for time, but derived tasks (T0009, T0010, etc.) are stopped. - It is easier to get started once early on, even with prototyping, so that it is easier to follow up.
.
--
.
1.### "remind before deadline" routine to reduce last minute rush . - A system of "notification one week before the deadline" through GitHub Actions, etc. - Tasks like 3/9 LT and 3/15 tax returns are dealt with in advance by dividing them into smaller groups.
2.### Start a small integrated test of information channels . - Put together only the most important chats/emails first → Effectiveness verification → Expansion - Do not try to integrate everything at once, but proceed in stages.
3.### periodically block time for document-based tasks . - Example: once a week, devote one or two hours only to documenting and organizing. - Steadily advance technical documentation and role definitions, which are often postponed.
4.### Simplified prototype precedence for dependent tasks . - MVID analysis → Even if it can be done roughly by manual or partial scripting, it should be shaped early to move the followings.
--
Summary
Highest priority: 3/9 LT materials (T0067) / 3/15 Tax returns (T0109)
Next steps: Information channel support measures (overlapping integration of T0113/T0116), AI instance role definition (T0012), MVID dataset analysis (T0008)
Improvement points: Automate pre-deadline reminders, regular block time for document-based tasks, and move dependent tasks early with simple prototypes.
First, let's concentrate on finishing tomorrow's LT materials, and then continue with the tax returns all at once.
next Weekly journal 2025-03-16~2025-03-24
This page is auto-translated from /nishio/週記2025-03-08~2025-03-16 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.